Self-construal and students’ math self-concept, anxiety ...€construal and learning 1...

Title Self-construal and students’ math self-concept, anxiety and achievement:

An examination of achievement goals as mediators Author(s) Wenshu Luo, David Hogan, Liang See Tan, Berinderjeet Kaur, Pak Tee Ng,

and Melvin Chan Source Asian Journal of Social Psychology, 17(3), 184–195.

http://dx.doi.org/10.1111/ajsp.12058 Published by Wiley This is the peer reviewed version of the following article: Luo, W., Hogan, D., Tan, L. S., Kaur, B., Ng, P. T., & Chan, M. (2014). Self-construal and students’ math self-concept, anxiety and achievement: An examination of achievement goals as mediators. Asian Journal of Social Psychology, 17(3), 184–195. http://dx.doi.org/10.1111/ajsp.12058, which has been published in final form at http://dx.doi.org/10.1111/ajsp.12058. This article may be used for non-commercial purposes in accordance with Wiley Terms and Conditions for Self-Archiving.

http://dx.doi.org/10.1111/ajsp.12058

http://olabout.wiley.com/WileyCDA/Section/id-828039.html#terms

Self‐construal and learning

Running head: SELF-CONSTRUAL and LEARNING

Self-construal and students’ math self-concept, anxiety and achievement: An examination

of achievement goals as mediators

Wenshu Luo*, David Hogan, Liang See Tan, Berinderjeet Kaur, Pak Tee Ng, and Melvin Chan

Nanyang Technological University

*Correspondence should be addressed to Dr. Wenshu Luo, Policy and Leadership Studies

Academic Group, National Institute of Education, 1 Nanyang Walk, Singapore 637616; Email:

[email protected]; Tel: (65) 6790-3235; Fax: (65) 6316-4787.


1

Self-construal and students’ math self-concept, anxiety and achievement: An examination

of achievement goals as mediators

Abstract

This study examines the role of self-construal in student learning by testing a mediation

model: through math achievement goals, self-construal predicts math self-concept and anxiety,

which further predict math achievement. A sample of 1196 students from 104 Singapore

secondary classes took a survey and a math achievement test. The results from multi-group

structural equation modeling supported measurement invariance and equal path coefficients in

the mediation model between boys and girls. Interdependent self-construal predicted positively

mastery approach and avoidance goals, through which interdependent self-construal had a

positive total indirect effect on math anxiety. Independent self-construal predicted positively

mastery approach, performance approach and performance avoidance goals, and through the two

approach goals, high independent self-construal was associated with high math self-concept.

Overall, self-construal was not associated with math achievement. The findings enhance our

understanding of achievement motivation from a sociocultural perspective and help explain East

Asian students’ relatively higher anxiety and lower self-concept in comparison with their

Western counterparts reported in international studies.

Keywords: self-construal, achievement goals, achievement, self-concept, anxiety


2

Introduction

Cross-cultural research has found that people have different views of the self in relation to

others that are specific to particular cultures (Kitayama, Markus, Matsumoto, & Norasakkunkit,

1997; Markus & Kitayama, 1991). In general, people in Western individualistic cultures seek to

maintain their independence from others by attending to the self and expressing their unique

inner attributes (independent self-construal). In contrast, people in collectivistic cultures, such as

in East Asian countries, tend to include important others in their self-definition and emphasize

harmonious interdependence with each other (interdependent self-construal). Although cultural

contexts typically promote the development of one or the other self-construal more strongly,

most researchers agree that to some extent both self-views exist in all societies and individuals

vary in the extent to which they construe the self in the culturally mandated way (e.g., Markus &

Kitayama, 1991; Oyserman & Lee, 2008; Singelis, 1994). Researchers have proposed that

different self-views have important implications for cognition, emotion and motivation (Markus

& Kitayama, 1991). With respect to motivation, it is assumed that people with high

interdependent self-construal are more socially motivated to attain relations with significant

others and improve interconnectedness, while those with high independent self-construal have

more personal goals to differentiate themselves from others and enhance positive self-views. For

example, it was reported that individualism of Singapore undergraduate students predicted their

status and achievement motivation, but not their affiliation motivation (Brutus & Greguras,

2008), and German college students with high interdependent self-construal valued social goals

more than individuals goals (van Horen, Pohlmann, Koeppen, & Hannover, 2008).

The different emphasis on social goals relative to personal goals given by students with

independent and interdependent self-construals has implications for achievement motivation.


3

Although students with both independent and interdependent self-construals might have some

social purposes for academic achievement, their social goals have different meanings for

achievement motivation and behaviors (Urdan & Maehr, 1995). In general, studies have found

that some social goals, such as pleasing teachers and bringing honor to parents, function more

like internalized or intrinsic motivation in collectivistic cultures but more like externally

regulated or extrinsic motivation in individualistic cultures. For example, Lepper, Corpus, and

Iyengar (2005) reported that there was a positive correlation between intrinsic motivation and the

desire for pleasing teachers for Asian students, but a negative correlation for Caucasians.

Similarly, some researchers proposed that some social goals for academic achievement are more

akin to mastery goals for students with interdependent self-construal than for those with

independent construal (Cheng & Lam, 2013; Urdan & Maehr, 1995). In particular, Cheng and

Lam (2013) found that social goals, such as to be a good son/daughter or to prove teachers’

teaching quality, were associated with higher willingness to improve after failure and lower

report of avoidance behavior for Hong Kong students with interdependent self-construal than

their counterparts with independent self-construal.

It is reasonable that compared with people in Western cultures, people from East Asian

cultures are more ready to internalize the expectations of important others because important

others are part of their self-definition (Markus & Kitayama, 1991). As a result, these

expectations become more internal regulations (Ryan & Deci, 2000), which lead to mastery-

oriented motivation and learning behaviors. This is consistent with the self-improvement

tendency of East Asian people, who tend to make effort to improve in the areas where they have

failed to do well. In contrast, Americans have a general self-enhancement tendency, that is, they

tend to value or actively seek positive self-relevant information to maintain or enhance a positive


4

self-view (Heine et al., 2001; Heine, Lehman, Markus, & Kitayama, 1999; Kitayama, et al.,

1997). For example, Heine et al. (2001) reported that Japanese who failed on a task persisted

working on the same type of task longer than those who succeeded, while North Americans who

succeeded persisted longer on the same type of task.

In this study, we hypothesize that self-construal has an impact on student learning through

the mediation of one important motivation variable, achievement goal orientations, and we test

this hypothesis with secondary students in a modernized East Asian country, Singapore. Through

expanding the research of self-construal to education, this study not only contributes to our

knowledge about the role of self-construal in the motivational processes of academic

achievement, but also enhances our understanding of student motivation and learning from a

sociocultural perspective. In addition, results from large-scale international studies, such as

Program for International Student Assessment (PISA), have shown that compared to their

Western counterparts, East Asian students tend to have a combination of high performance, low

self-concept and high anxiety (Lee, 2009; Wilkins, 2004). By investigating the role of self-

construal in student motivation and learning, this study also helps understand these puzzling

findings.

Self-construal, achievement goals and learning

Over the last three decades, the theory of achievement goals has been one of the most

important frameworks to understand students’ motivation and learning. In recent years, both

mastery and performance goals distinguished in earlier research have been bifurcated by the

approach-avoidance distinction, which leads to the 2 × 2 achievement goal framework (Elliot &

McGregor, 2001). Individuals approaching an activity with mastery approach goals make effort

to develop their knowledge and skills, while students approaching an activity with mastery


5

avoidance goals are concerned about task-based or intrapersonal incompetence. Individuals with

performance approach goals strive to outperform others or to demonstrate their competence

relative to others, while students pursue performance avoidance goals to avoid lower

performance than others or unfavorable judgments of ability.

Numerous studies have examined the relationships of achievement goals to students’

competence beliefs, anxiety and performance. Competence beliefs were measured as either self-

efficacy or self-concept, or both. Although, moderate to high correlations are frequently reported

between these two variables (e.g., Marsh, Dowson, Pietsch, & Walker, 2004; Yeung, Craven, &

Kaur, 2012), we include self-concept in this study because it entails a social comparison process

(Marsh, Trautwein, & Ludtke, 2008). Such a social comparison process is theoretically relevant

to the social psychological nature of self-construal, which is defined in relation to others.

Adaptive relationships between mastery approach goals and students’ competence beliefs,

anxiety and performance have been reported in the literature (e.g., Luo, Paris, Hogan, & Luo,

2011; Murayama & Elliot, 2009; Yeung, et al., 2012), although a positive relationship between

mastery approach goals and achievement was not always found (for a review, see Hulleman,

Schrager, Bodmann, & Harackiewicz, 2010). Recent research on mastery avoidance goals

reported that this dimension was positively related to anxiety (Putwain, Woods, & Symes, 2010;

Sideridis, 2008) and negatively associated with perceived competence (Cury, Elliot, Da Fonseca,

& Moller, 2006; Putwain, et al., 2010). In addition, mastery avoidance goals were either

unrelated or deleterious to achievement (Elliot & Murayama, 2008; Liem, Martin, Porter, &

Colmar, 2012; Van Yperen, Elliot, & Anseel, 2009).

A maladaptive learning profile has been associated with performance avoidance goals, but

results are less consistent with performance approach goals. Performance avoidance goals have


6

been related to low perception of competency, high anxiety, and low grades (e.g., Luo, Paris, et

al., 2011; Pajares, Britner, & Valiante, 2000; Putwain, et al., 2010; Urdan, Ryan, Anderman, &

Gheen, 2002). Recent reviews showed that the mixed findings about performance approach goal

orientation are related to its definition and measurement (Hulleman, et al., 2010; Senko,

Hulleman, & Harackiewicz, 2011). When measured as normative comparison (doing better than

others), performance approach goals showed a moderate average correlation with performance

avoidance goals ( = .34), and a positive average relationship with performance ( = .14). In

contrast, when measured as competence demonstration, performance approach goals showed a

high average correlation with performance avoidance goals ( = .71), and an average negative

relationship with performance ( = -.14). Both types of measures have been positively related to

self-concept or self-efficacy (e.g., Linnenbrink, 2005; Murayama & Elliot, 2009; Pajares, et al.,

2000; Putwain, et al., 2010). However, performance approach goals measured as normative

comparison were negatively related or unrelated to test anxiety, negative affect and avoidance

behaviors (e.g., Elliot & McGregor, 2001; Howell & Buro, 2009; Howell & Watson, 2007; Shih,

2005), while performance approach goals measured as competence demonstration were

positively related or unrelated to these variables (e.g., Linnenbrink, 2005; Luo, Paris, et al., 2011;

Midgley, Kaplan, & Middleton, 2001; Pajares, et al., 2000).

Although self-construal has important implications for achievement motivation, little

research has directly examined the relationship between self-construal and students’ achievement

goals. One study examined the predictive relationship between self-construal and achievement

goals with Singapore secondary students in learning English (Luo, Hogan, & Paris, 2011). They

reported that both independent and interdependent self-construal predicted positively mastery

approach goals, while only interdependent self-construal predicted positively mastery avoidance


7

goals. In addition, only independent self-construal predicted positively performance approach

and avoidance goals. The mastery orientation of students with interdependent self-construal

might explain their self-improvement tendency to work hard to improve in the areas they lack

mastery.

The finding about self-construal and performance goals was partly supported by a more

recent study (Cheng & Lam, 2013) with Chinese students in Hong Kong, which reported that

independent self-construal was positively related to performance approach goal orientation,

while interdependent self-construal was unrelated to it. The positive relationship between

independent self-construal and performance goals might reflect the self-enhancement tendency

of independent students. In particular, achieving performance goals by demonstrating high

performance or avoiding demonstration of low competence relative to others might be a way to

maintain or enhance a positive self-regard. The non-relationship between interdependent self-

construal and performance goals might reflect the tradeoff among different types of social

purposes for academic achievement of interdependent students. On one hand, interdependent

students might want to achieve social approval through demonstrating competence to teachers

and classmates, because academic achievement is highly valued in Asian contexts. This is

supported by the finding that Asian students’ social approval motive is positively associated with

both mastery and performance goals (Chang & Wong, 2008; Liem, et al., 2012). On the other

hand, they might also want to achieve interpersonal harmony with peers, while self-enhancement

through competence demonstration in social comparison might be detrimental to this social goal

(Heine, et al., 1999).


8

Purposes and hypotheses of this study

In this study, we aim to examine the relationships between self-construal and achievement

goals in a different sample of Singapore secondary students in their math study. Furthermore, we

aim to test a mediation model of achievement goals in the relationship between self-construal

and three learning variables, math self-concept, anxiety and achievement. Students’ self-concept

and emotions play a very important role in self-regulated learning and achievement. Researchers

have found reciprocal relationships between academic self-concept and achievement (Guay,

Marsh, & Boivin, 2003; Marsh & Yeung, 1997). In addition, anxiety has been generally related

to low performance (e.g., Linnenbrink, 2005; Pajares, et al., 2000; Pekrun, Elliot, & Maier, 2009;

Zeidner, 1998).

The hypothesized mediation model is given in Figure 1. More specifically, we hypothesized

that interdependent self-construal would predict positively mastery approach and avoidance

goals, and independent self-construal would predict positively mastery approach, performance

approach, and performance avoidance goals. In addition, although we expected that the tension

between seeking different social goals would lead to non-relationship between interdependent

self-construal and performance goals, we would like to test this assumption by comparing nested

models.

Based on the literature on achievement goals, we predicted that mastery approach goals, in

turn, would predict positively math self-concept and negatively math anxiety, and the contrary

for mastery avoidance goals. Since in this study performance goals were measured as students’

concern about competence demonstration, we expected that performance approach and

avoidance goals would show a high correlation with each other and similar correlational patterns

with other variables. However, after controlling for each other, performance approach goals


9

would have an adaptive relationship, and performance avoidance goals would have a

maladaptive relationship with math self-concept and anxiety.

In addition, we would control for previous achievement in this model. We expected that

previous achievement would positively predict mastery approach goals, but negatively predict

the other goals. In addition, previous achievement would predict math self-concept and math

achievement positively and math anxiety negatively. After controlling for previous achievement,

we expected that math self-concept would predict math achievement positively and math anxiety

would predict math achievement negatively.

Method

Participants and procedure

This study was part of a large-scale project that examined classroom practices in Singapore

schools. Half of the Second 3 (Grade 9) students in the same class were group administered an

online survey and then an online math achievement test in their computer laboratories with an

interval of 1-3 weeks. The survey was given in English, the medium of instruction in Singapore.

The average time for both the survey and assessment was about 40 minutes. There were 115

students who did not take the assessment after the survey, and they were not significantly

different from the other 1196 students with complete data in their math scores in Primary School

Leaving Examination (t (1309) = -1.63, p =.10). The 115 cases were excluded in the analysis.

The 1196 students from 104 classes included 616 (51.5%) boys with an average age of 15.44

years (SD = .60). Ethnic composition was Chinese (892, 74.6%), Malay (178, 14.9%), Indian (68,

5.7%), and others (58, 4.8%).


10

Measures

Self-construal. Based on meta-analyses at regional and country levels, Oyserman et al.

(2002) proposed that the principal way to operationalize individualism is to measure the extent to

which personal uniqueness and independence is valued, and that the principal way to

operationalize collectivism is to measure the extent to which duty to in-group and group

harmony is valued. Based on this definition, we measured self-construal by selecting or adapting

items in existing scales. We selected five items with relatively high loadings (over .44) on

interdependent self-construal in Singelis’ (1994) Self-Construal Scale to measure this dimension.

Sample items include, “My happiness depends on the happiness of those around me,” and “It is

important for me to maintain harmony with my group.” We measured independent self-construal

using two items adapted from Singelis (1994) with high loadings (over .5) on independent self-

construal, two items adapted from Shulruf, Hattie, and Dixon (2007) with high loadings (.60) on

uniqueness, and one item adapted from Oyserman (1993). Sample items are, “I consider myself

as a unique person separate from others,” and “I see myself as a very independent person.” The

response categories ranged from 1 (strongly disagree) to 5 (strongly agree). The internal

reliabilities for interdependent and independent self-construal were .72 and .77, respectively.

Achievement goals. The scales employed to measure mastery approach goals (3 items),

performance approach goals (3 items), and performance avoidance goals (3 items) were adapted

from the Patterns of Adaptive Learning Scale (Midgley et al., 2000). This instrument was chosen

in the project for comparison with an earlier data set using the same instrument. However, since

it doesn’t measure mastery avoidance goals, we assessed this dimension using 3 items adapted

from the Achievement Goal Questionnaire (Elliot & McGregor, 2001). The internal reliabilities

for mastery approach (e.g., “An important reason I do my math work is that I like to learn new


11

things”), mastery avoidance (e.g., “I’m afraid that I may not understand the content of my math

class thoroughly”), performance approach (e.g., “I want to show my classmates in my math class

that I am smart”), and performance avoidance goals (e.g., “It is very important that I do not look

stupid in front of my classmates during my math class”) were .84, .78, .85, and .82, respectively.

Math self-concept. The scale of math self-concept (4 items) adapted from PISA (2003)

measured students’ perception of how good they are in learning math, such as “I have always

believed that math is one of my best subjects.” The internal reliability of this scale was .89.

Math anxiety. Four items adapted from PISA (2003) were used to measure students’ anxiety

experienced in learning math, such as “I often worry that it will be difficult for me in math

classes.” The internal reliability of this scale was .87.

Math achievement. To measure math achievement, an online multiple-choice test with 28

items was constructed by a small group of researchers and experienced teachers with reference to

the curriculum. The test included questions assessing knowing, applying, and reasoning abilities

in four content areas, including Number, Algebra, Measurement and Geometry, as well as

Statistics and Probability. The internal reliability of the test was .86. In addition, students were

also asked to report their Primary School Leaving Examination (PSLE) scores in math taken

three years earlier. The PSLE math scores ranged from 1 to 7, with higher scores indicating

higher performance in this study.

Statistical analyses

Before testing the mediation model using structural equation modeling analysis, we

conducted some preliminary analyses to examine the nature of the data. First, following Baron

and Kenny’s (1986) recommendations, we checked whether the predictor, mediator and outcome

variables were correlated. We also recognized that a significant correlation between the predictor


12

and the outcome variable is not a prerequisite for a mediation effect, such as when one of the

mediators works as a suppressor (MacKinnon, Krull, & Lockwood, 2000). Second, due to the

hierarchical nature of the data, we calculated intra-class correlations (ICCs) to decompose the

variance at student and class levels. This helped us decide whether the class level variances

should be considered in the modeling. Third, we conducted multi-group confirmatory factor

analyses to test the measurement model and measurement equivalence between girls and boys.

To reduce complexity, the 28 items in the achievement test were grouped into four composite

indicators according to the four content domains. For PSLE math scores, since there was only

one indicator, we constrained its residual variance to be zero to facilitate model identification.

Before testing the complete mediation model in Figure 1, we tested a more complex mediation

model, where we allowed the direct paths from self-construal to math self-concept and anxiety to

be estimated together with the mediating effects. Following Baron and Kenny’s (1986), this is an

important step to decide whether achievement goals function as mediators. We then compared

this model with the model in Figure 1 to test our assumption that achievement goals fully

mediate the relationship between self-construal and student learning.

Results

As shown in Table 1, independent and interdependent self-construals were moderately

correlated with each other (r = .50), which supports the finding in previous studies that the two

types of self-construal coexist in Asian cultures (Cheng & Lam, 2013; Luo, Hogan, et al., 2011).

Low or moderate correlations were found between the two types of mastery goals and between

mastery goals and performance goals (rs = .11 to .32). Consistent with the review of studies

measuring performance goals as competence demonstration (Hulleman, et al., 2010), there was a

high correlation (r = .76) between performance approach and avoidance goals. In addition, the


13

two types of self-construal were related to the four types of achievement goals (rs = .13 to .24),

which were associated with all the three outcome variables (rs = -.12 to .49). In addition, both

types of self-construal were correlated positively with math self-concept and anxiety (rs = .07

to .14), but not math achievement.

As shown in Table 1, only the two achievement measures had ICC larger than .10. All the

self-construal and achievement goal measures had ICCs smaller than .05, indicating less than 5%

of the total variances due to variations at class level. As a result, we decided to test the

measurement and mediation model only at student level in Mplus 6.11, but we would take class

as a cluster variable in order to have more accurate standard errors (Krull & MacKinnon, 2001).

Based on suggestions in the literature (Sharma, Durvasula, & Ployhart, 2012; Vandenberg &

Lance, 2000), we compared three nested models in sequence to test measurement equivalence

between boys and girls: Model 1 with configural invariance (the same factor pattern), Model 2

with metric equivalence (the same factor loadings), and Model 3 with both metric and scalar

equivalence (the same factor loadings and intercepts of indicators). The goodness-of-fit indexes

of the three models are shown in Table 2. Since chi-square statistic is sensitive to sample size,

researchers have suggested using the difference in other goodness-of-fit statistics to compare

nested models. For example, if the difference in comparative fit index (CFI) between two nested

models is no larger than .01, the more restricted model should not be rejected (Cheung &

Rensvold, 2002). Comparing the three models in Table 2, we can see that the most parsimonious

Model 3 was supported, indicating measurement equivalence between boys and girls. The

standardized factor loadings are shown in Table 3, along with the explained variances in

observed indicators as local fit indexes (Ray & Zajacova, 2012). In general, the local fit of the


14

model was supported, but the data also suggest the measure of self-construal can be improved in

future studies.

With both metric and scalar equivalence met, gender differences in the latent variables were

tested in Model 3 and the effect sizes (D) were calculated (Sharma, et al., 2012). Gender

differences were found in mastery avoidance goals (D = -.15, p =.02), performance approach

goals (D = .33, p =.00), performance avoidance goals (D = .28, p =.00), math self-concept (D

= .29, p =.00), and math achievement (D = -.23, p =.02). Girls were higher in mastery avoidance

goals, while boys had higher performance approach and avoidance goals. This is generally

consistent with previous research that found girls tended to be mastery-oriented and boys were

more performance oriented (e.g., Elliot & Church, 1997; Luo, Hogan, et al., 2011). In addition,

girls were also higher in current math achievement but lower in math self-concept, which is

consistent with previous findings (e.g., Herbovich, Sirsch, & Felinger, 2004; Marsh, 1989).

Before testing the mediation model, we tested the invariance of the factor variance and

covariance matrix between girls and boys (Model 4). If the fit of Model 4 is not worse than that

of Model 3, any structural model among the latent variables will be invariant across the two

groups (Vandenberg & Lance, 2000). However, as shown in Table 2 this was not supported and

thus in the following steps we constrained only the regression coefficients to be equal. Before

testing the hypothesized mediation model (Model 5), we first tested a more complex mediation

model (5.0) which allowed the direct paths from self-construal to math self-concept and anxiety.

Comparing Model 5.0 with Model 3, we can see that the equivalence of regression coefficients in

Model 5.0 was supported. In addition, achievement goals significantly predicted math concept

and anxiety after controlling for self-construal, supporting the mediational role of achievement

goals. Then, we compared Model 5 with Model 5.0 and found Model 5 was superior to Model


15

5.0 in which all the direct paths from self-construal to math self-concept and anxiety were non-

significant. Thus, the hypothesized complete mediation model was supported. In Model 5,

however, we found that the high correlation between performance approach and avoidance goals

led to large standard errors and non-significant path coefficients from the two performance goals

to math self-concept and anxiety. Large standard errors typically lead to unstable parameter

estimates in the model (Marsh, et al., 2004). Therefore, based on theoretical analysis, in Model 6

we kept only the paths from performance approach goals to math self-concept and from

performance avoidance goals to math anxiety. As shown in Table 2, Model 6 was supported.

For comparison purpose, we also conducted Model 6.1, in which we kept the paths from

performance approach goals to math anxiety and from performance avoidance goals to math self-

concept. As shown in chi-square, Model 6 was superior to Model 6.1.We then examined the path

coefficients from interdependent self-construal to the two performance goals, and found that both

path coefficients were not significant. In Model 7, we further removed these two paths and found

Model 7 was superior to Model 6. The equal unstandardized path coefficients across boys and

girls in Model 7 are given in Figure 2.

As shown in Figure 2, the hypothesized mediation model was supported. Interdependent

self-construal predicted positively mastery approach and avoidance goals and independent self-

construal predicted positively mastery approach, performance approach, and performance

avoidance goals. Mastery approach goals further predicted positively math self-concept and

negatively math anxiety. The opposite was found with mastery avoidance goals. Performance

approach goals further predicted positively math self-concept and performance avoidance goals

predicted positively math anxiety. The indirect effects from interdependent self-construal to math

self-concept and anxiety through mastery approach and avoidance goals, from independent self-


16

construal to math self-concept through mastery and performance approach goals, and from

independent self-construal to math anxiety through performance avoidance goals were all

significant (p < .05). After controlling for previous achievement, math self-concept predicted

math achievement positively and math anxiety predicted math achievement negatively.

As shown in Table 4, there were significant total indirect effects from independent self-

construal to math self-concept (β = .19), from interdependent self-construal to math anxiety (β

= .11), and from previous achievement to current math achievement (β = .05). In addition, the

total indirect effects from the four achievement goals to math achievement were all significant:

from mastery achievement goals (β =.37), from mastery avoidance goals (β = -.44), from

performance approach goals (β = .04), and from performance avoidance goals (β = -.08).

Discussion

Self-construal was assumed to have important motivational implications (Kitayama, et al.,

1997; Markus & Kitayama, 1991). In this study, based on theoretical analysis and empirical

findings, we proposed that self-construal is associated with one important motivational variable,

achievement goals, which mediate the relationship of self-construal to three learning variables:

math self-concept, anxiety and achievement. The proposed mediation model was generally

supported in the multi-group structural equation modeling analysis. The findings of this study

advance our understanding of achievement motivation from a sociocultural perspective and help

explain cross-cultural differences in students’ motivation and emotion.

Consistent with our hypothesis, interdependent self-construal only uniquely predicted

mastery approach and avoidance goals. Since students with interdependent self-construal tend to

endorse mastery goals in their study, this helps explain the important finding in the literature that

people from Confucian cultures incline towards self-improvement through mastery oriented


17

learning behaviors in response to failure (Heine, et al., 2001; Heine, et al., 1999; Kitayama, et al.,

1997). The mastery orientation of students with interdependent self-construal might be related to

the values in Confucian cultures. The Chinese culture has strong emphasis on the importance of

both academic achievement and the role of effort in academic achievement (Hau & Salili, 1991;

Li, 2002). These values are readily internalized by Chinese children as internal regulation due to

a sense of interdependence and responsibility for important others (Markus & Kitayama, 1991).

For example, Chinese children regard effort as the most important attribution for academic

performance (Hau & Salili, 1991; Luo, Hogan, Yeung, Sheng, & Aye, 2013). As a result, the

cultural emphasis on effort may prompt an incremental view of ability and mastery-oriented

learning behaviors (Hau & Salili, 1991; Salili & Hau, 1994). However, this does not mean that

the cultural emphasis on academic achievement and effort is totally adaptive for students with

interdependent self-construal. The unique contribution of interdependent self-construal to

mastery avoidance goals indicates that interdependent students also tend to focus their regulatory

attention on failing to meet internal standards of competence. This mastery avoidance tendency

might be an important factor that pressures Chinese children to study for long hours (Salili, Chiu,

& Lai, 2001) and explains their sensitiveness to negative information (Heine, et al., 2001;

Kitayama, et al., 1997) and relatively high anxiety as found in international studies (Lee, 2009;

Wilkins, 2004)

One important difference in the findings between interdependent and independent self-

construal is that only independent self-construal uniquely predicted performance goals. People

with independent self-construal attend to the expression of their unique internal attributes and the

enhancement of positive self-regards (Kitayama, et al., 1997; Markus & Kitayama, 1991). Thus,

it is possible that independent students pursue their study not only to master knowledge and


18

skills but also to maintain or enhance self-views by demonstrating high performance or avoiding

demonstration of low competence. The non-significant relationship between interdependent self-

construal and performance goals might be due to the tradeoff among various social goals of

interdependent students. As argued by Urdan and Maehr (1995), the various social goals, such as

social approval goals, social responsibility goals, social status goals, prosocial goals, and social

affiliation goals, may have distinct meanings for motivation and achievement. We argue that

students with interdependent self-construal have various social goals, which might have opposite

impacts on the adoption of performance goals. For example, social approval goals might promote

the adoption of performance goals, but social affiliation goals could lead to reduced performance

goals. Future studies can measure different types of social goals and examine their roles in the

relationship between self-construal and achievement goals.

The findings of this study provide a possible way to explain East Asian students’ higher

anxiety and lower self-concept in comparison with their Western counterparts in international

studies. That is, the differences might be at least partly related to one sociocultural indicator,

self-construal. More specifically, Eastern Asian students might have higher interdependent self-

construal, which, through the mediation of mastery avoidance goals, leads to high math anxiety.

In addition, Western students might have high independent self-construal, which, through the

mediation of performance approach goals, leads to high math self-concept. The relationship

between self-construal and students’ math anxiety and self-concept is also consistent with the

finding in social psychology that collectivistic culture or interdependent self-construal was

related to high social anxiety and low self-esteem (e.g., Heine, et al., 1999; Okazaki, 1997; Xie,

Leong, & Feng, 2008). It should be noted that in this study we did not find that self-construal

was a factor to explain differences in achievement between students from Western and East


19

Asian cultures. The higher performance of East Asian students might be related to other factors,

such as learning strategies and teacher education.

In general, the findings of this study enhance our knowledge of the role of self-construal in

the motivational processes of academic achievement. That is, self-views that are related to

particular cultures have important implications for learning through orienting people towards the

adoption of different achievement goals. The findings also improve our understanding of student

achievement motivation from a sociocultural perspective. In particular, the findings imply that to

promote adaptive learning we should take into account students’ self-views. For students with

interdependent self-construal, we should focus on reducing their concern about not meeting

internal standards of excellence and its distractions of learning process (Sideridis, 2008), while

for students with independent self-construal, we should focus on reducing their concern about

demonstration of incompetence in social comparison and its negative effects on learning.

Some limitations should be considered when readers interpret the findings in this study.

First, although we hypothesized a mediation model of achievement goals in the relationship

between self-construal and learning, the presumed causal relationships in the model cannot be

tested in this study due to the cross-sectional design. Second, although the mediation model was

generally supported in the study, the role of self-construal in student motivation and learning was

relatively small since only about 4 -12% of the variances in the four achievement goal

orientations were explained by the predictor variables. In addition, the hypothesized mediation

model was modified to some extent to derive the final model to fit the data, particularly due to

the high correlation between performance approach and avoidance goals. Therefore, more

studies should be conducted to test whether the final model can be replicated. Third, in this

study we assessed mastery avoidance goals by using items adapted from the original


20

Achievement Goal Questionnaire (AGQ, Elliot & McGregor, 2001). These items have explicit

reference to affective content, which might to some extent confound the relationship between

mastery avoidance goals and the affective variables in this study. Future studies can test the

findings by using more recently developed instruments, such as Elliot, Murayama, and Pekrun’s

(2011) recently developed 3 × 2 Achievement Goal Questionnaire. In addition, recent reviews

have distinguished two basic components of performance goals, normative comparison and

competence demonstration, and found they have different impacts on student learning (Hulleman,

et al., 2010; Senko, et al., 2011). In this study, performance goals were measured as competence

demonstration, and thus future studies can examine whether there is a different relationship

between self-construal and the normative component of performance goals. Furthermore, this

study was conducted in Singapore and the findings may not fully explain cross-cultural

differences because self-construal is just one of the many indicators of culture. Interested

researchers can examine the role of self-construal together with other sociocultural factors in

students’ achievement motivation and learning in cross-cultural studies.


21

References

Baron, R. M., & Kenny, D. A. (1986). The moderator-mediator variable distinction in social

psychological research: Conceptual, strategic, and statistical considerations. Journal of

Personality and Social Psychology, 51, 1173-1182.

Brutus, S., & Greguras, G. J. (2008). Self-construals, motivation and feedback-seeking behaviors.

International Journal of Selection and Assessment, 16(3), 282-291.

Chang, W. C., & Wong, K. (2008). Socially oriented achievement goals of Chinese university

students in Singapore: Structure and relationships with achievement motives, goals and

affective outcomes. International Journal of Psychology, 43(5), 880-885.

Cheng, R. W., & Lam, S. F. (2013). The interaction between social goals and self-construal on

achievement motivation. Contemporary Educational Psychology, 38, 136-148.

Cheung, G. W., & Rensvold, R. B. (2002). Evaluating goodness-of-fit indexes for testing

measurement invariance. Structural Equation Modeling, 9(2), 233-255.

Cury, F., Elliot, A. J., Da Fonseca, D., & Moller, A. C. (2006). The social-cognitive model of

achievement motivation and the 2 x 2 achievement goal framework. Journal of

Personality and Social Psychology, 90, 666-679.

Elliot, A. J., & Church, M. (1997). A hierarchical model of approach and avoidance achievement

motivation. Journal of Personality and Social Psychology, 72, 218-232.

Elliot, A. J., & McGregor, H. A. (2001). A 2 x 2 achievement goal framework. Journal of

Personality and Social Psychology, 80(3), 501-519.

Elliot, A. J., & Murayama, K. (2008). On the measurement of achievement goals: Critique,

illustration, and application. Journal of Educational Psychology, 100(3), 613-628.


22

Elliot, A. J., Murayama, K., & Pekrun, R. (2011). A 3 x 2 achievement goal model. Journal of

Educational Psychology, 103(3), 632-648.

Guay, F., Marsh, H. W., & Boivin, M. (2003). Academic self-concept and academic achievement:

Developmental perspectives on their causal ordering. Journal of Educational Psychology,

95, 124-136.

Hau, K. T., & Salili, F. (1991). Structure and semantic differential placement of specific causes:

Academic causal attributions by Chinese students in Hong Kong. International Journal of

Psychology, 26, 175-193.

Heine, S. J., Kitayama, S., Lehman, D., Takata, T., Ide, E., Leung, C., et al. (2001). Divergent

consequences of success and failure in Japan and North America: An investigation of

self-improving motivations and malleable selves. Journal of Personality and Social


Heine, S. J., Lehman, D. R., Markus, H. R., & Kitayama, S. (1999). Is there a universal need for

positive self-regard? Psychological Review, 106, 766-794.

Herbovich, A., Sirsch, U., & Felinger, M. (2004). Gender differences in the self-concept of

preadolescent children. School Psychology International, 25(2), 207-222.

Howell, A. J., & Buro, K. (2009). Implicit theories, achievement goals, and procrastination: A

mediational analysis. Learning and Individual Differences, 19, 151-154.

Howell, A. J., & Watson, D. (2007). Procrastination: Association with achievement goal

orientation and learning strategies. Personality and Individual Differences, 43(1), 167-

178.


23

Hulleman, C. S., Schrager, S. M., Bodmann, S., & Harackiewicz, J. M. (2010). A meta-analytic

review of achievement goal measures: Different labels for the same constructs or

different constructs with similar labels. Psychological Bulletin, 136(3), 422-449.

Kitayama, S., Markus, H. R., Matsumoto, H., & Norasakkunkit, V. (1997). Individual and

collective processes in the construction of the self: self-enhancement in the United States

and self-criticism in Japan. Journal of Personality and Social Psychology, 72, 1245-1267.

Krull, J. L., & MacKinnon, D. P. (2001). Multilevel modeling of individual and group-level

mediated effects. Multivariate Behavioral Research, 36, 249-277.

Lee, J. (2009). Universals and specifics of math self-concept, math self-efficacy, and math

anxiety across 41 PISA 2003 participating countries. Learning and Individual Differences,

19, 355-365.

Lepper, M. R., Corpus, J. H., & Iyengar, S. S. (2005). Intrinsic and extrinsic motivational

orientations in the classroom: Age differences and academic correlates. Journal of

Educational Psychology, 97(2), 184-196.

Li, J. (2002). A cultural model of learning: Chinese heart and mind for wanting to learn. Journal

of Cross-Cultural Psychology, 33, 246-267.

Liem, A. G. D., Martin, A. J., Porter, A. L., & Colmar, S. (2012). Sociocultural antecedents of

academic motivation and achievement: Role of values and achievement motives in

achievement goals and academic performance. Asian Journal of Social Psychology, 15,

1-13.

Linnenbrink, E. A. (2005). The dilemma of performance-approach goals: The use of multiple

goal contexts to promote students' motivation and learning. Journal of Educational



24

Luo, W., Hogan, D., & Paris, S. G. (2011). Predicting Singapore students' achievement goals in

their English study: Self-construal and classroom goal structure. Learning and Individual

Differences, 21, 526-535.

Luo, W., Hogan, D., Yeung, A. S., Sheng, Y. Z., & Aye, K. M. (2013). Attributional beliefs of

Singapore students: Relations to self-construal, competence, and achievement goals.

Educational Psychology: An International Journal of Experimental Educational

Psychology, DOI: 10.1080/01443410.01442013.01785056.

Luo, W., Paris, S. G., Hogan, D., & Luo, Z. (2011). Do performance goals promote learning? A

pattern analysis of Singapore students' achievement goals. Contemporary Educational


MacKinnon, D. P., Krull, J. L., & Lockwood, C. M. (2000). Equivalence of the mediation,

confounding and suppression effect. Prevention Science, 1, 173-181.

Markus, H. R., & Kitayama, S. (1991). Culture and self: Implications for cognition, emotion, and

motivation. Psychological Review, 98, 224-253.

Marsh, H. W. (1989). Age and sex effects in multiple dimensions of self-concept:

Preadolescence to early adulthood. Journal of Educational Psychology, 81, 417-430.

Marsh, H. W., Dowson, M., Pietsch, J., & Walker, R. (2004). Why multicollinearity matters: A

reexamination of relations between self-efficacy, self-concept, and achievement. Journal

of Educational Psychology, 96(3), 518-522.

Marsh, H. W., Trautwein, U., & Ludtke, M. (2008). Social comparison and Big-Fish-Little-Pond

effects on self-concept and other self-belief cosntructs: Role of generalized and specific

others. Journal of Educational Psychology, 100(3), 510-524.


25

Marsh, H. W., & Yeung, A. S. (1997). Causal effects of academic self-concept on academic

achievement: Structural equlation models of longitudinal data. Journal of Educational


Midgley, C., Kaplan, A., & Middleton, M. (2001). Performance approach goals: good for what,

for whom, under what circumstances, and at what cost? Journal of Educational


Midgley, C., Maehr, M. L., Hruda, L. Z., Anderman, E., Anderman, L., & Freeman, K. E. (2000).

Manual for the Patterns of Adaptive Learning Scales (PALS). Ann Arbor: University of

Michigan.

Murayama, K., & Elliot, A. J. (2009). The joint influence of personal achievement goals and

classroom goal structures on achievement relevant outcomes. Journal of Educational

Psychology, 101(2), 432-447.

Okazaki, S. (1997). Sources of ethnic differences between Asian American and While American

college students on measures of depression and social anxiety. Journal of Abnormal


Oyserman, D. (1993). The lens of personhood: Viewing the self and others in a multicultural

society. Journal of Personality and Social Psychology, 65(5), 993-1009.

Oyserman, D., Coon, H. M., & Kemmelmeier, M. (2002). Rethinking individualism and

collectivism: Evaluation of theoretical assumptions and meta-analyses. Psychological

Bulletin, 128, 3-72.

Oyserman, D., & Lee, W. S. (2008). Does culture influence what and how we think? Effects of

priming individualism and collectivism. Psychological Bulletin, 134, 311-342.


26

Pajares, F., Britner, S. L., & Valiante, G. (2000). Relation between achievement goals and self-

beliefs of middle school students in writing and science. Contemporary Educational


Pekrun, R., Elliot, A. J., & Maier, M. A. (2009). Achievement goals and achievement emotions:

Testing a model of their joint relations with academic performance. Journal of

Educational Psychology, 1, 115-135.

Putwain, D. W., Woods, K. A., & Symes, W. (2010). Personal and situational predictors of test

anxiety of students in post-compulsary education. British Journal of Educational


Ray, T., & Zajacova, A. (2012). On latent change model choice in longitudinal studies.

Structural Equation Modeling, 19(4), 580-592.

Ryan, R. M., & Deci, E. L. (2000). Self-determination theory and the facilitation of intrinsic

motivation, social development, and well-being. American Psychologist, 55(1), 68-78.

Salili, F., Chiu, C., & Lai, S. (2001). The influence of culture and context on students'

motivational orientation and performance. In F. Salili, C. Chiu & Y. Hong (Eds.), Student

motivation: The culture and context of learning. New York: Kluwer Academic/Plenum.

Salili, F., & Hau, K. T. (1994). The effect of teachers' evaluative feedback on Chinese students'

perception of ability: A cultural and situational analysis. Educational Studies, 20, 223-

236.

Senko, C., Hulleman, C. S., & Harackiewicz, J. M. (2011). Achievement theory at the crossroads:

Old contraversies, current challenges, and new directions. Educational Psychologist,

46(1), 26-47.


27

Sharma, S., Durvasula, S., & Ployhart, R. E. (2012). The analysis of mean differences using

mean and covariance structure analysis: Effect size estimation ad error rates.

Organizational Research Methods, 15(1), 75-102.

Shih, S. S. (2005). Taiwanese sixth graders' achievement goals and their motivation, strategy use,

and grades: An examination of the multiple goals perspective. The Elementary School

Journal, 106(1), 39-58.

Shulruf, B., Hattie, J. A., & Dixon, R. (2007). Development of a new measurement tool for

individualism and collectivism. Journal of Psychoeducational Assessment, 25(4), 385-

401.

Sideridis, G. D. (2008). The regulation of affect, anxiety, and stressful arousal from adopting

mastery-avoidance goal orientations. Stress and Health, 24, 55-69.

Singelis, T. M. (1994). The measurement of independent and interdependent self-construals.

Personality and Social Psychology Bulletin, 20, 580-591.

Urdan, T. C., & Maehr, M. L. (1995). Beyond a two-goal theory of motivation and achievement:

A case for social goals. Review of Educational Research, 65(3), 213-243.

Urdan, T. C., Ryan, A. M., Anderman, E. M., & Gheen, M. H. (2002). Goals, goal structures,

and avoidance behaviors. In C. Midgley (Ed.), Goals, goal structures, and patterns of

adaptive learning (pp. 55-83). Mahwah, NJ: Erlbaum.

van Horen, F., Pohlmann, C., Koeppen, K., & Hannover, B. (2008). Importance of personal goals

in people with independent versus interdependent selves. Social Psychology, 39, 213-221.

Van Yperen, N. W., Elliot, A. J., & Anseel, F. (2009). The influence of mastery-avoidance goals

on performance improvement. European Journal of Social Psychology, 39, 932-943.


28

Vandenberg, R. J., & Lance, C. E. (2000). A review and synthesis of the measurement invariance

literature: Suggestions, practices and recommendations for organizational research.

Organizational Research Methods, 3(1), 4-70.

Wilkins, J. M. (2004). Mathematics and science self-concept: An international investigation. The

Journal of Experimental Education, 72(4), 331-346.

Xie, D., Leong, F. T. L., & Feng, S. (2008). Culture specific personality correlates of anxiety

among Chinese and Caucasian college students. Asian Journal of Social Psychology, 11,

163-174.

Yeung, A. S., Craven, R., & Kaur, G. (2012). Mastery goal, value and self-concept: What do

they predict? Educational Research, 54(4), 469-482.

Zeidner, M. (1998). Test anxiety: The state of the art. New York: Plenum.


29

Table 1

Descriptive Statistics, Correlation Coefficients, and Intra-Class Correlations (ICCs)

Mean SD 1) 2) 3) 4) 5) 6) 7) 8) 9) 10) ICCs

1) Interdependent self-construal 3.71 0.58 1 .50* .22* .16* .13* .14* .11* .11* .00 .04 .02

2) Independent self-construal 3.76 0.60 1 .24* .14* .18* .14* .14* .07* .02 .04 .01

3) Mastery approach 3.55 0.77 1 .11* .32* .23* .49* -.14* .17* .08* .04

4) Mastery avoidance 3.42 0.83 1 .21* .27* -.18* .44* -.12* -.00 .03

5) Performance approach 2.86 0.91 1 .76* .22* .15* -.14* -.05 .03

6) Performance avoidance 2.85 0.91 1 .12* .25* -.20* -.09* .04

7) Math self-concept 3.22 0.90 1 -.24* .26* .15* .10

8) Math anxiety 2.82 0.90 1 -.29* -.08* .09

9) Current math achievement 14.42 6.18 1 .32* .56

10) PSLE math 4.91 1.60 1 .36

Note. * p < .01.


30

Table 2

Goodness-of-Fit Indexes for Measurement and Structural Models

χ2 (df) CFI TLI RMSEA SRMR

Model 1 (configural invariance) 1666.15 (1032) .957 .951 .032 .042

Model 2 (metric equivalence) 1694.38 (1057) .957 .952 .032 .044

Model 3 (scalar equivalence) 1074.67 (1082) .953 .949 .033 .044

Model 4 (Factor variance and covariance

invariance)

2505.16 (1102) .905 .898 .046 .080

Model 5.0 1850.96 (1124) .951 .948 .033 .051

Model 5 1857.39 (1128) .951 .948 .033 .051

Model 6 1857.53 (1130) .951 .948 .033 .052

Model 6.1 1862.63 (1130) .951 .948 .033 .052

Model 7 1860.55 (1132) .951 .948 .033 .052


31

Table 3

Standardized Factor Loadings in Model 3 and Explained Variances (R2) in Dependent Indicators

Variable PSLEM ITSC IDSC MAPG MAVG PAPG PAVG MSC MANX MACH se R2

PSLEM 1.0/1.0 .00/.00 1.0/1.0 ITSC1 .66/.63 .03/.04 .44/.39 ITSC2 .59/.52 .04/.05 .35/.28 ITSC3 .65/.56 .04/.05 .42/.31 ITSC4 .56/.53 .04/.05 .31/.28 ITSC5 .59/.54 .04/.04 .34/.29 IDSC1 .72/.71 .03/.03 .52/.50 IDSC2 .60/.48 .03/.03 .37/.24 IDSC3 .70/.69 .03/.03 .49/.47 IDSC4 .60/.54 .03/.03 .36/.29 IDSC5 .70/.69 .03/.03 .50/.48 MAPG1 .79/.81 .03/.02 .63/.66 MAPG2 .81/.81 .02/.02 .65/.66 MAPG3 .78/.80 .03/.02 .61/.64 MAVG1 .64/.68 .03/.03 .41/.46 MAVG2 .76/.82 .03/.03 .58/.67 MAVG3 .73/.78 .03/.03 .54/.61 PAPG1 .76/.77 .02/.03 .58/.59 PAPG2 .81/.80 .02/.02 .65/.64 PAPG3 .82/.85 .02/.02 .68/.72 PAVG1 .79/.82 .02/.02 .63/.67 PAVG2 .82/.82 .02/.02 .68/.67 PAVG3 .71/.72 .03/.03 .50/.52 MSC1 .84/.81 .02/.02 .71/.65 MSC2 .85/.81 .02/.02 .73/.66 MSC3 .77/.75 .02/.02 .59/.56 MSC4 .86/.82 .02/.02 .73/.67 MANX1 .79/.81 .02/.02 .62/.65 MANX2 .77/.79 .03/.02 .59/.63 MANX3 .78/.79 .02/.02 .61/.63 MANX4 .77/.81 .03/.02 .59/.65 MACH1 .71/.76 .02/.02 .50/.57 MACH2 .81/.82 .02/.02 .65/.68 MACH3 .75/.76 .03/.02 .56/.58 MACH4 .65/.67 .03/.03 .42/.44

Note. The values before the slash are for boys and after the slash are for girls.

PSLEM = PSLE math, ITSC = Interdependent self-construal, IDSC = Independent self-construal, MAPG = Mastery approach, MAVG= Mastery avoidance, PAPG = Performance approach, PAVG = Performance avoidance, MSC = Math self-concept, MANX = Math anxiety, and MACH = Math achievement.


32

Table 4

Total Indirect Effects and Total Effects on the Three Learning Variables

Total indirect effect Total effect On math self-concept Interdependent self-construal .07 (.10) .07 (.10) Independent self-construal .19* (.07) .19* (.07) Previous achievement .02 (.02) .10* (.03)

On math anxiety Interdependent self-construal .11* (.06) .11* (.06) Independent self-construal -.02 (.04) -.02 (.04) Previous achievement -.03 (.02) -.05* (.02)

On math achievement Interdependent self-construal -.03 (.05) -.03 (.05) Independent self-construal .06 (.04) .06 (.04) Previous achievement .05* (.01) .31* (.06) Mastery approach .37* (.06) .37* (.06) Mastery avoidance -.44* (.06) -.44* (.06) Performance approach .04* (.02) .04* (.02) Performance avoidance -.08* (.02) -.08* (.02) Note. p < .05. Standard errors are in parentheses.


33

Figure Captions

Figure 1.The hypothesized mediation model.

Note. The dashed lines indicate relationships to be tested by comparing models.

Figure 2.The final mediation model (Model 7).

Note. Only significant path coefficients and their standard errors are reported. The values in the

parentheses are percentage explained variances in latent mediator and outcome variables for

boys and girls, respectively.

Figure 1

34

Figure 2

35

Self-construal and students’ math self-concept, anxiety ...€construal and learning 1...

Documents

Transcript of Self-construal and students’ math self-concept, anxiety ...€construal and learning 1...